Even if he had a save in every one of those games, he’s pitched one inning per game, and the team delivered a lead to him over the previous 8 innings. If you accept the fact that any pitcher pitching any inning has some positive value to his team (with the obvious negatives subtracted for giving up runs, hits, and walks), then you must accept that a pitcher who pitches only 62 innings must pitch phenomenally well to overcome the huge contribution even a mediocre starting pitcher makes via additional innings. So Gil Meche has pitched 129 innings, about twice as much as Mo. Even a losing pitcher adds value because he eats up innings that better pitchers don’t have to waste their energy on. So Mo is useless without a heck of a lot of other people making very positive contributions to the Yankees.

You have to accept, just on his limited role alone, that it would be very unlikely for any closer to be the most valuable player on a team. I’m not saying it’s impossible, but it’s very unlikely. If anyone could be, it’s Mo, but before declaring something that’s highly unlikely, you should probably dig deeper and see what you find. And looking at highlight reels of the past 10 years doesn’t count as digging deeper.

Or you can just continue to yell at anyone who says something surprising to you.

Notice no one said “Meche is as good as Mariano.” Just “as valuable” because of the number of innings he’s contributed to his team.

This is no knock on Mo. He’s a special talent.

That’s ridiculous. Just because a stat doesn’t give you 100% certainty doesn’t mean it’s worthless.

WAR does not attempt to discuss the true talent of a player. It looks at the actual data of actual games played, compiles the data on all of the events that player was involved in, compares his success/failure, chances, etc. etc. against a replacement level player and assigns a number of wins above replacement based on a clear, well defined set of metrics for every subset of the game.

It’s not made out of whole cloth.

Taking defensive metrics specifically, yes of course you can’t look at one individual play and say “it was in his zone, so he should have gotten it.” But you can certainly look at all shortstops in the history of baseball, see that the average shortstop gets to X% of balls hit into a certain zone, see that Ozzie Smith gets to a much high percentage than average as well as many outside his zone, and conclude that Ozzie Smith is a heck of a short stop. Zone ratings do not attempt to grapple with individual plays. They put them all into a bucket and see what comes out.

These metrics show that Derek Jeter was a subpar short stop for the past several years, but to the lay fan, he may just think all those hits up the middle were just that. Hits. Jeter has always been fundamentally sound, so he “looks” like a good fielder. He is a good fielder. He just had limited range. But this year he’s dramatically improved going to his left. I’ve seen plays he’s made this year that I know he never would have gotten to last year. And UZR reflects that.

The question for any stat’s usefulness is, “does it improve our understanding of the game and who is succeeding at the game?” In both UZR and WAR’s cases, the answer is undeniably yes. It matters not whether they prove beyond a shadow of a doubt who’s the best defensive player in baseball. No one ever gave them that much credit. They are data points, pieces of evidence a critical thinking person uses to piece together the entire story.

Defensive statistics are just as valid as offensive and pitching statistics. We dont judge pitchers on how they look on the mound. We judge them on how they convert pitched balls into outs. We dont just hitters on how their swing looks on the jumbotron. We judge them on how they put balls into play. The same goes for defense. It doesnt matter how often a guy makes a routine play look spectacular. Defensive stats tell us how often they get to balls and make outs. They arent perfect, but neither is OB or opponents OB.

Sabermetrics are statistics that are valid within their frame of reference. They arent predictive and they arent fact.